The effect of substrate molecular mobility on surface induced immune complement activation and blood plasma coagulation

ARTICLE IN PRESS

Biomaterials 25 (2004) 4581–4590

The effect of substrate molecular mobility on surface induced immune complement activation and blood plasma coagulation Mattias Berglin*, Marcus Andersson, Anders Sellborn, Hans Elwing Cell and Molecular Biology, Interface Biophysics, Goteborg University, Box 462, SE-405 30, Sweden . Received 5 September 2003; accepted 22 November 2003

Abstract Changing the length of the alkyl ester side chain in poly(alkyl methacrylates) provides a unique opportunity to systematically vary the mobility of the polymer chains, or in other words vary the glass transition temperature (Tg ), without greatly affect the solid surface energy (gs ) of the polymer. A series of poly(alkyl methacrylate) coatings was therefore analysed with regard to the human immune complement (IC) activation and the surface associated blood plasma coagulation cascade (CC) properties. For the IC and CC measurements we used a quartz crystal microbalance (QCM) where we modified the chemistry of the sensor surface by applying 10–30 nm thick poly(alkyl methacrylate) coatings. The surface energy was calculated from water contact angles and small differences between the coatings were observed. The surface chemistry of the coatings, as determined with X-ray photoelectron spectroscopy (XPS), showed no deviation from expected compositions. Tapping mode atomic force microscopy (TM-AFM) measurements revealed that all coatings displayed similar morphology and the roughness was in the range of 0.7–0.9 nm. Increased polymer mobility correlated with a decrease in IC activation, measured as a decreased C3c deposition at the surface. The surface induced CC, measured as fibrin clot formation at the surface, was different between the different coatings but no correlation with molecular mobility was observed. Thus, the molecular mobility of the polymer chains had a major effect on both the IC and the CC and it seems that different aspects of the chemistry of the solid surface regulate activation of the IC and the CC. r 2003 Elsevier Ltd. All rights reserved. Keywords: Protein adsorption; Immune complement system; Plasma coagulation; Quartz crystal microbalance

1. Introduction It is well known that biomaterial surfaces in contact with blood both activate the immune complement (IC) system [1–4] and trigger the blood plasma coagulation cascade (CC) [5–9], resulting in unwanted effects such as acute inflammatory reaction (IC related) and thrombus formation (CC related). The effect of surface properties such as interfacial energetics and surface morphology on both the IC and the CC has been thoroughly investigated. However, a direct correlation between surface properties and the biological responses has not yet been established. The lack of a fundamental model testifies to the complexity of this problem. Furthermore, to be able to construct a biological model where blood or tissues are in contact with biomaterials you need to system*Corresponding author. Tel.: +46-31-7732577; fax: +46-317732599. E-mail address: [email protected] (M. Berglin). 0142-9612/$ - see front matter r 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2003.11.050

atically investigate all possible variables. One such variable may be the polymer chain mobility. It is very reasonable to believe that polymer chain mobility would influence the interfacial reactions of solute molecules with the polymer surface. The mobility of polymer molecules is expressed via a very commonly measured bulk property of polymers, the glass transition temperature (Tg ). It should be stressed that in most systems it is difficult to vary the Tg without varying the surface chemistry or the surface morphology. However, the family of alkyl methacrylate polymers with different length alkyl ester side chain provides a means of being able to vary Tg without greatly altering surface wetting characteristics. In this study, we used a series of poly(alkyl methacrylates) with a wide range of different glass transition temperatures with the aim to study if the molecular mobility of the substrate might have an effect on the surface induced IC activation and the blood plasma CC. The traditional methods to measure IC activation, for example the enzyme-linked immunosorbent assay

ARTICLE IN PRESS 4582

M. Berglin et al. / Biomaterials 25 (2004) 4581–4590

(ELISA), are based on measurements of fluid phase components such as the C3a, which is formed due to cleavage of C3 [10]. However, methods measuring fluid phase components neglect the fraction of activated complement proteins bound to the surface [1,2]. This may either result in underestimating the total amount of activated complement proteins produced, or lead to an erroneous conclusion due to the possibility that the surface bound components may have a different reactivity than or different functionality than the fluid components. One method that is able to determine the amount of the surface bound IC components is the enzyme-linked immunoflow assay (ELIFA) developed by Black and Sefton [11]. They showed that the membrane attack complex (C5b-9) could be desorbed from the polyvinyl alcohol (PVA) surface with sodium dodecyl sulphate (SDS) and analysed together with the C5b-9 present in the bulk. Furthermore, the amount of surface bound complement proteins on material surfaces have been evaluated with an enzyme immunoassay based on monoclonal antibodies specific for neoepitopes exposed in complement activation products [12]. We have used a different technique to quantify the surface bound complement proteins based on a quartz crystal microbalance (QCM) [13]. The shift in resonance frequency (Df ) of the sensor crystal when human antiC3c is added is interpreted as an arbitrary measure of IC activation. The major advantage with the QCM technique besides measuring surface bound IC components is the possibility to study the adsorption kinetics and calculation of the total mass of the adsorbed layer [14]. Furthermore, there is no need for optical clarity of the test solutions, which is an advantage over the optical methods such as ellipsometry. With the same argument, using fluid phase components of the contact activated blood CC, such as the b-FXIIa or the thrombin– antithrombin III complex (TAT), as an indicator of surface thrombogenicity might lead to incorrect conclusions. In the case of CC measurements we use the QCM apparatus to study the fibrin clot deposition at the sensor surface. From the measurements we determine the time until deposition starts (onset time), the deposition rate, and finally, when the reaction stops and no more deposition of fibrin can be detected at the surface (coagulation time) [15].

2. Experimental

Poly(octadecyl methacrylate)

POMA

Poly(hexyl methacrylate)

PHMA

Poly(isobutyl methacrylate), PIBMA

Poly(lauryl methacrylate)

PLMA

Poly(butyl methacrylate)

PBMA

Fig. 1. Repeating units of the poly(alkyl methacrylate) coatings used in this study.

was provided already dissolved in toluene and was diluted with toluene to a final concentration of 0.5% (w/v). Poly(hexyl methacrylate) (PHMA) (Scientific polymer Products, USA, cas# 25087-17-6) and poly(octadecyl methacrylate) (POMA) (Polysciences Inc., USA, cas# 25639-21-8) were provided already dissolved in toluene and were diluted with toluene to a final concentration of 0.5% (w/v). The poly(butyl methacrylate) (PBMA) was made in-house by radical polymerisation. The butyl methacrylate monomer was purchased from Fluka, Sweden (cas# 96-05-9). The monomer was purified from the inhibitor, hydroquinone-monomethylether, by adsorption chromatography using ICN alumina N Super I from ICN Biomedicals, Germany. The polymer was synthesised by bulk radical polymerisation at 60
C for 12 h with 2,20 -azobisisobutyronnitrile (AIBN) from Fluka, Sweden (cas# 97-88-1) at a concentration of 2.0 mg ml1 monomer. The polymer was dissolved in toluene and precipitated in cold methanol before being re-dissolved in toluene at a final concentration of 0.5% (w/v). The repeating units of the polymers used in this study are shown in Fig. 1. Polymeric coatings were spin-coated on quartz crystals by applying 50 ml of the polymer solutions at 2000 rpm for 1 min.

2.1. Polymeric substrates 2.2. Characterisation of pristine coatings Poly(isobutyl methacrylate) (PIBMA), and poly(lauryl methacrylate) (PLMA) was provided by SigmaAldrich (cas# 9011-15-8 and cas# 25719-52-2, respectively). PIBMA was dissolved in toluene (Merck, Sweden) to a final concentration of 0.5% (w/v). PLMA

The glass transition temperatures (Tg ) provided by the manufacturers were used except in the case of the inhouse synthesised PBMA polymer. The Tg of PBMA was determined by differential scanning calorimetry

ARTICLE IN PRESS M. Berglin et al. / Biomaterials 25 (2004) 4581–4590

(DSC) on a Perkin Elmer, Pyris-1. A heating and cooling rate of 10
C min1 from 14
C to 80
C was used for the analysis. The Tg was determined as the midpoint of the heat capacity change. X-ray photoelectron spectroscopy (XPS) was done using a Physical Electronics Quantum 2000 scanning ESCA microprobe. The X-ray source was an Al Ka anode with a beam size of 100 mm operating at 20W/ 15 kW. All measurements were made at a 45
take-off angle, and the area analysed was 0.5  0.5 mm. Binding energy shifts were resolved using the peak for hydrocarbon (C1 s) at 285.0 eV as a reference. A 3 eV electron flood gun was used for charge neutralisation during analysis. Sensitivity factors determined for this instrument were used to calculate elemental compositions. Advancing and receding contact angles were determined by the Wilhelmy plate method [16] at 20
C. The instrument used was a DCA-322 from Cahn. The instrument was operating at 100 mm s1. Data were evaluated with WinDCA version 1.03 (Cahn, WI, USA). Wettability measurements were made using de-ionised water (g ¼ 72:8 mJ/m2). The surface energies (gs ) of the poly(alkyl methacrylate) coatings were calculated using an equation-of-state approach formulated by Wulf et al. [17], Eq. (1). rffiffiffiffiffiffiffi 2 g cos y ¼ 1 þ 2 sv ebðglv gsv Þ ; ð1Þ glv where b is a constant determined to 0.0001247 (m2/mJ)2, glv is the surface tension of the probe liquid and cos y is the advancing or receding water contact angle. Cover glass slides covered with polymer were used in these experiments. The surface morphology was examined with a Digital Instruments NanoScope III, atomic force microscope fitted with a NanoScope IIIA controller and a Dimension 3000 large sample type G scanner. The coatings were analysed in tapping mode with standard silicon tips. Analyses were made at set point ratio (Asp =A0 ) of 0.7–0.8, where Asp is the set point amplitude and A0 is the free oscillation amplitude of the cantilever. At least an area of 5  5 mm was used in the calculation of the root mean square roughness (RMS). 2.3. Preparation of human serum, blood plasma and buffer solutions A more detailed description of preparation of human sera, blood plasma and buffer solutions can be found elsewhere [13,15]. In summary, fresh, citrate anticoagulated, human blood plasma from three healthy donors was obtained from the blood central of . Sahlgrenska University hospital, Goteborg. The plasma was pooled and immediately frozen (80
C) until further use. It should be noted that the procedure used for the plasma preparation might not remove the

4583

presence of cell fragments totally. To completely avoid these remnants of cells the plasma should be ultracentrifuged or filtered. This means that the coagulation responses might not necessarily be entirely constituted of fibrin polymerisation. It could be that the different poly(alkyl methacrylate) coatings attract these cell fragments to different extents. For sera preparation, fresh and citrate anti-coagulated human blood plasma from healthy donors was mixed with CaCl2 (0.25 m). The plasma was then incubated for 30 min in 37
C. The clotted plasma was centrifuged (Universal 16R, 5200 rpm, 3000g) for 30 min in 37
C. The clot was discarded and the serum was immediately frozen at 80
C until further use. Heat inactivated sera (30 min in 56
C) were used for control measurements. Rabbit anti human C3c was obtained from Dako, Denmark. The anti human C3c was diluted 1:50 in Veronal buffered saline (VBS), (0.15 m NaCl, 1,8 mm Na-5-5diethylbarbiturate, 3.1 mm 5-5-diethylbarbituric acid, 0.5 mm NaN3, pH 7.4), supplemented with 0.5 mm MgCl2 and 0.15 mm CaCl2 (VBS++). 2.4. Surface induced immune complement and plasma coagulation studies For immune complement and coagulation studies we used a method developed in our lab that we have named Acoustic Blood Biointerface Analysis or ‘‘ABBA’’. The ABBA methodology is based on the surface sensitive quartz crystal microbalance (QCM) apparatus [14,18]. The quartz crystal microbalance used in this study was a . Q-Sense D300 (Q-Sense AB, Goteborg, Sweden) with a temperature controlled fluid cell. In summary, the QCM technique is an acoustic method where a quartz crystal is set in lateral resonance oscillation with a predefined frequency (f ). Simultaneous frequency and dissipation (d) measurements are made by periodically switching on and off the AC-voltage over the crystal. The decay signal is recorded and fitted to an exponentially damped sinusoidal curve. The adsorbed amount (g cm2) can be calculated from Eq. (2) by the frequency shift using the Sauerbrey equation [19], provided that the mass is evenly distributed, does not slip on the electrode and is sufficiently rigid and/or thin to have negligible internal friction. Gold plated quartz crystals (Q-Sense, Sweden) were cleaned according to the procedure published elsewhere [13]. Polymeric coatings were spin-coated on the crystals by applying 50 ml of the polymer solutions at 2000 rpm for 1 min. Thickness (d; cm) of the spin-coated polymers were calculated from the change in resonance frequency (Df ) of the crystals according to Eq. (2). Cf is the mass sensitive constant (e.g. 17.7  109 g cm2 Hz1 at f ¼ 15 MHz), nr is the shear wave number, rpolymer is the density (g cm3) of the polymer and mcoating is the

ARTICLE IN PRESS 4584

M. Berglin et al. / Biomaterials 25 (2004) 4581–4590

mass per unit area (g cm2) of the polymer coating. Cf  Df ¼ rpolymer dcoating ¼ mcoating : nr

ð2Þ

The IC part of the ABBA experiments were done using the following two step procedure: (1) establishing a stable baseline with pure VBS++ buffer for at least 5 min; (2) sera (diluted 1:20 with VBS++ buffer) was added to the sample chamber and the protein adsorption was observed during 30 min; (3) washing with pure VBS++ buffer for 5 min; (4) antibodies (anti human C3c) was added and the adsorption was studied for 20 min; (5) washing with pure VBS++ buffer. The shift in resonance frequency when the anti human C3c was added was taken as an arbitrary measure of IC activation. C3c is soluble degradation product originating from iC3b. A fraction of C3c re-adsorbs and it is this fraction of re-adsorbed C3c that we measure with the ABBA method. C3c is not incorporated in any other complement protein complexes such as the C3 convertase and the membrane attack complex (MAC) C5b9. Moreover, C3c undergoes slow degradation and is therefore suitable to analyse when measuring the activation of the immune complement system. The total mass of bound anti-C3c was calculated with Eq. (2). The presented value is the average of at least four measurements. Control experiments were carried out with quartz crystals with a layer of pre-adsorbed IgG on the sensor surface, which have been shown to demonstrate IC activation [20] and with heat-inactivated sera, 56
C for 30 min (no activation of the IC) [13]. At least three control experiments for each surface were carried out. For plasma clot formation studies 100 ml, citrated plasma was incubated on the coated quartz crystal sensor. After about 2 min a fresh solution of plasma (90 ml) and CaCl2 (10 ml 0.25 m) was added and the fibrin clot deposition, i.e. the time until fibrin deposition starts (Onset time) or lag phase, and the deposition rate (Df =min), which was calculated at the inflection point, were studied. Moreover, from the measurement we also determined the total coagulation time, or when no more fibrin was deposited at the surface. At least two measurements for each surface were carried out. Titanium (Ti), or more correctly Titaniumdioxide (TiO2), has been shown to be highly thrombogenic [9] and was therefore used as positive control. Sensor surfaces with a thin layer of sputtered Titanium, which in air rapidly oxidises to Titaniumdioxid (TiO2), were . obtained from Q-Sense AB (Goteborg, Sweden). The Ti surfaces were prepared according to the following procedure. The sensor surfaces were cleaned and exposed to Argon (Ar) plasma for 5 min in the sputter chamber. The Ti coatings were produced in a Leybold DC-magnetron Z600 sputtering chamber. A 300 nm thick Ti layer was sputtered on the surfaces. The root mean square roughness (RMS) was determined to be

3 nm with TM-AFM measurements. The Ti surfaces were cleaned in 2% SDS (w/v) solution followed by thoroughly rinsing in de-ionised water and 10 min treatment in an UV/ozone chamber prior to the CC measurements. A heparinised surface (kindly provided by Corline systems AB, Uppsala, Sweden) was used as negative control [21]. The heparin-coated surfaces were prepared by irreversible adsorption of a macromolecular conjugate composed of 70 Heparin molecules covalently linked to a 50 kDa polymer carrier. The surface concentration of Heparin was approximately 0.5 mg cm2. At least five control experiments for each control surface were performed.

3. Results and discussion 3.1. Polymer characterisations The spin-coating process resulted in coatings ranging from 10 up to 26 nm in thickness as calculated from Eq. (2), Table 1. The table presents the average of at least three coatings. The variation in thickness was in the range of 10–15%. The appearance of the coated sensor surfaces was satisfying and no de-wetting of the polymer film or other surface imperfections could be detected with light microscopy. The general polymer characteristics together with the surface properties of the pristine polymers are summarised in Table 1. As expected, the Tg decreases with increased number of carbons in the alkyl ester side chain, with the highest Tg for PIBMA (66
C) and the lowest for POMA (100
C). Thus, at 22
C the polymer chain mobility of PIBMA will be constrained with no or small chain movements. In contrast, the polymer chains of POMA, PLMA, PHMA and PBMA will be in constant motion. The XPS results show no deviation from expected stoichiometric composition, Table 1. Hence, no inadvertent surface contaminants due to the preparation of the coated quartz crystals could be detected. The average roughness (RMS), as determined by TM-AFM, ranged from 0.7 to 0.9 nm for all coatings as can be seen in Table 1. Consequently, differences in biological response are not a result of different surface morphologies. Furthermore, no cracks and other surface imperfections that could affect the biological response were found with TM-AFM. The advancing and receding water contact angles of the coatings are shown in Fig. 2. The advancing contact angle showed a slight increase with an increase in number of carbons in the side chain as has been found in several other studies [17,22]. Interestingly, the advancing contact angle of PBMA and PIBMA were comparable but the receding contact angle of PIBMA was higher. This can be explained by the bulkier isobutyl group is hiding the hydrophilic ester group from the solid–water

ARTICLE IN PRESS M. Berglin et al. / Biomaterials 25 (2004) 4581–4590

4585

Table 1 General polymer characteristics; shown is the polymer abbreviation with the number of carbons in alkyl side chain, glass transition temperature (Tg ) and coating thickness (calculated from Eq. (2)) Polymer (# of carbons Glass transition in alkyl side chain) temperature (Tg ) (
C)

Coating thickness (nm)

C=O ratio (calculated)

C=O ratio (measured)

Roughness (RMS) (nm)

Surface energy gs ; (advancing/receding) (mJ/m2)

PIBMA (4) PBMA (4) PHMA (6) PLMA (12) POMA (18)

26 11 10 12 13

4.0 4.0 5.0 8.0 11.0

4.2 4.1 5.2 8.6 10.9

0.88 0.70 0.70 0.70 0.76

28.3/36.5 28.3/40.3 23.1/38.6 19.1/40.5 17.6/42.8

66 17 6 70 100

The surface chemistry was analysed with XPS and the carbon to oxygen elemental ratio was calculated based on the repeating unit of the polymer and compared with measured. Surface roughness (root mean square roughness, RMS) was determined with TM-AFM and the surface energy was calculated from water contact angles using Eq. (1). Note that the surface energy was calculated both from advancing and receding contact angles.

120

-30 PIBMA (inactivated sera) PLMA PIBMA

-40

Shift in frequency, ∆ff (Hz)

Contact angle (degrees)

110

100

90

80

PIBMA PBMA

70

Serum exposure Wash Anti-C3c

-50

Wash

-60

∆f

-70

-80

60 -90

50

0

0

5

10

15

20

Number of carbons in alkyl side chain

Fig. 2. Advancing (filled diamonds) and receding (open circles) water contact angles of the poly(alkyl methacrylate) coatings as a function of number of carbons in alkyl ester side chain.

interface and thereby increase the receding contact angle. The somewhat lower receding contact angles for the PLMA and POMA compared to PBMA and PIBMA can be explained by higher molecular mobility, allowing ester groups to orient towards the interface during the measurement. The solid surface energy of the coatings were calculated from the advancing and receding contact angles using an equation-of-state approach published by Wulf et al. [17], Table 1. The surface energy calculated from advancing contact angles ranged from 28.3 mJ/m2 (PIBMA and PBMA) down to 17.6 mJ/m2 (POMA). On the other hand, it can be argued that the receding contact angle should be used due to the fact that all the experiments were carried out in aqueous environment. Using the receding contact angle when calculating the solid surface energy significantly increases the surface energy of the coatings. For example, the surface energy of PIBMA was calculated to 36.5 mJ/m2 and the surface energy of

10

20

30

40

50

60

7 70

80

Time (min)

Fig. 3. The shift in frequency (Df ) plotted as a function of time during serum exposure followed by a wash with VBS++ buffer and addition of human anti-C3c. Representative measurements of PLMA (open circles) and PIBMA (filled triangles) are shown. Also shown is the significant reduction of anti-C3c binding obtained with heat-inactivated serum using PIBMA as substrate (crossed lines).

POMA was calculated to 42.8 mJ/m2. The difference between the coatings was still minor as can be seen in Table 1 and should not affect the interpretation of the IC and CC results. 3.2. Immune complement (IC) activation Two representative experiments (PIBMA and PLMA) are shown in Fig. 3. The experiments were carried out as follows: when a stable baseline was reached human serum was added and the decrease in frequency due to protein adsorption was studied. The adsorption of serum proteins to the rather hydrophobic coatings was very fast. However, it can be noted that the kinetics was somewhat faster for PIBMA, where steady state was reached already after a few minutes, Fig. 3. After 30 min of serum exposure no significant difference in the total mass of serum proteins among the coatings could

ARTICLE IN PRESS M. Berglin et al. / Biomaterials 25 (2004) 4581–4590

4586

be observed. Thus, after 30 min of serum exposure all polymeric surfaces accumulated the same mass of serum protein, which was calculated to 0.970.1 mg/cm2 using Eq. (2). This amount includes the water hydrodynamically coupled to the adsorbed protein film [23]. After a 5 min wash with VBS++ buffer the anti-C3c was added and the shift in frequency was studied. The difference in IC activation between the different polymer surfaces was apparent as can be seen in Fig. 4 and Table 2. Since the major difference between the polymers are the molecular mobility the results obtained in this study strongly indicates a correlation between molecular mobility and IC activation. The IC activation of the coatings can be divided into three groups; POMA demonstrated very low activation, slightly higher activity were achieved with PLMA, PHMA and PBMA (the middle group of polymers demonstrated statistically significant higher IC activity compared to POMA with p-values ranging from 0.002 and down when calculated with a student’s t-test, assuming equal variances). Finally, PIBMA demonstrated the highest activity. The PIBMA activity was significantly higher than POMA, PLMA and PHMA (po0:05) but not in the case of PBMA (p > 0:05). There was no significant difference between the polymers in the middle group consisting of PLMA, PHMA and PBMA (p > 0:05). We expected the IC activation of POMA, PLMA, PHMA and PBMA to be comparable due to the fact that the mobility of polymer chains is enhanced at surfaces and these four polymers should have a ‘‘surface’’ Tg well below the measurement temperature of 22
C. For example the surface Tg of atactic polypropylene has been calculated to 40
C, which is about 22
C lower than the bulk Tg [24]. Control experiments were carried out with heat-inactivated sera and a shift in frequency of

Immune complement activity (∆ Hz)

25

20

15

10

5 POMA PLMA Tg = -100 C Tg = -70 C

PHMA Tg = -6 C

PBMA Tg = 17 C

PIBMA Tg = 66 C

0

Fig. 4. The surface induced immune complement (IC) activation measured as shift in frequency (Df ; Hz) when the anti-C3c was added (7SE). The different polymers are placed on the x-axis with the lowest Tg to the left and the highest Tg to the right.

7 Hz was observed when anti-C3c was added. The shift in frequency with heat-inactivated sera was significantly different compared to all polymers except POMA with p-values ranging from 0.03 when comparing against PLMA, and down to 0.007 when comparing against PIBMA (t-test, assuming equal variance). It can be speculated that a rigid surface induces greater conformation changes in the adsorbed proteins. Conformational changes of protein secondary structure due to adsorption is generally accepted and has been shown in several studies [25–27]. However, none of the studies has investigated the effect of molecular mobility on the degree of conformational change. For example, adsorption of the complement protein C3 may result in a conformational change leading to the disruption of the thioester bond as suggested by Andersson et al. [28]. This ‘‘contact activation’’ of the C3 protein might initiate the alternative pathway of the IC activation. The degree of conformational change or the rate of the conformational change upon protein adsorption might be dependent on the molecular mobility of the solid surface. A rigid surface might cause a higher residual stress in the proteins upon adsorption compared to a more flexible surface. For example, when a drop of water is placed on a surface the forces between the water and the surface is causing the surface to deform to some extent [29] and it has been shown that the contact angle of water correlates to the molecular mobility of the surface [30]. Similar behaviour might occur when proteins interacts with a solid surface. The strength of the interactions between the protein and the surface and the mobility of both the protein and the surface are ultimately responsible for the final configuration of the adsorbed protein, and thereby the biological response as schematically shown for C3 adsorption and IC activation in Fig. 5. It is important to be aware of that the poly(alkyl methacrylate) coatings in this study accumulated significant amounts of serum proteins but we do not know the composition of this protein film. As comparison, the mobility of poly(ethylene glycol) or PEG chains are high but a PEG coating resists the accumulation of a protein film. Consequently, reduced complement activation has been observed with PEG coatings [31]. The protein resistance of PEG coatings has been associated with two main mechanisms e.g. ‘‘steric repulsion’’ and ‘‘hydration/water structuring’’ [32]. None of these mechanisms should be applicable in the case of the poly(alkyl methacrylate) coatings used in this study. Interestingly, the very low IC activity of POMA is remarkable and that result needs some further attention. It has been shown that the relatively long alkyl chains in POMA are able to crystallise. The movement of the alkyl chains will be constricted in the crystalline phase resulting in rigid segments in a very mobile surrounding [33,34]. It might be that this ‘‘composite’’ surface

ARTICLE IN PRESS M. Berglin et al. / Biomaterials 25 (2004) 4581–4590

4587

Fig. 5. A schematic model showing the surface induced contact activation of the complement system. On a rigid surface the conformational change of C3 activates the reaction ending up with the formation of C3c. On a surface with high polymer chain mobility no or low conformational change results in low IC activation. 100 Onset time (min)

Heparin POMA PLMA PHMA PBMA PIBMA Titanium

0

Delta f (Hz)

composed of mobile and rigid domains are responsible for the very low activity of POMA. The degree of ‘‘surface’’ crystallinity was not measured in this study and the effect of such composite surface can only be speculative. However, previous research suggests that crystallinity might have an effect on the biological response. In one study increased crystallinity of poly(lactic acid) (PLA) particles elicited a smaller inflammatory response in alveolar macrophages [35]. It can be speculated that the crystalline phase somehow might affect both the adsorption and the conformational change of the proteins. Worth bringing up in this context is a study by Kanazawa et al., where a series of poly(alkyl methacrylates) were studied according to their skin sensitising potential and it was found that POMA showed a decreased sensitising potential compared to, for example, PLMA [36]. It is known that a monolayer of IgG stimulates the IC system [20] and the IC activating properties of the poly(alkyl methacrylate) coatings were compared to an IgG coated surface. When an IgG coated surface was used as substrate the decrease in frequency was as high as 133 Hz [13]. Furthermore, polystyrene (PS) showed a decrease of 35 Hz [13]. It can be noted that the Tg of PS is around 100
C and PS can therefore be considered as a molecularly rigid surface. Comparison with poly(methyl methacrylate) (PMMA), that have a Tg of 105
C, was done. The somewhat higher surface energy of PMMA excluded this polymer from our series of poly(alkyl methacrylate) coatings. The surface energy of PMMA was determined to 38.7 mJ/m2 (calculated from advancing contact angles) and 48.5 mJ/m2 (calculated from receding contact angle). The IC activity of PMMA was

-100

Fibrin deposition Rate CCk (Hz min-1)

-200

-300

Coagulation time (min) -400 0

20

40

60

80

100

120

140

Time (min)

Fig. 6. Blood plasma coagulation measured as fibrin deposition rate. Representative experiments of PIBMA (filled triangles), PBMA (open diamonds), PHMA (open triangles), PLMA (open circles) and POMA (open squares) are shown. The control experiments with a Heparinised surface (non-thrombogenic) and a highly thrombogenic surface, i.e. titanium (Ti) are also shown (filled circles and filled squares, respectively).

determined to 1676 Hz (7SE, n ¼ 7). Thus, PMMA give rise to slightly less IC activation than for example PIBMA, but it should be stressed that the experimental error was rather high in these experiments. 3.3. Blood plasma coagulation cascade (CC) The initial lag phase or time until onset represents the time until thrombin is formed via the intrinsic pathway

ARTICLE IN PRESS M. Berglin et al. / Biomaterials 25 (2004) 4581–4590

4588

Table 2 Summary of the immune complement (IC) activity of the poly(alkyl methacrylate) coatings used in this study, displayed as the average shift (7standard error, SE) in frequency (Hz) when the anti C3c was added Polymer

Glass transition temperature (Tg ) (
C)

IC activity, Df

IC activity (%)

Amount of bound anti-C3c (ng/cm2)

PIBMA PBMA PHMA PLMA POMA

66 17 6 70 100

2272.7 1772.7 1471.9 1371.4 770.7

17 13 11 10 5

396747 295747 251735 236724 121712

The IC% is the IC activity normalized against an IgG coated surface. Finally, the mass (ng/cm2) of bound anti-C3c was calculated from Eq. (2). The average of at least 4 measurements is presented.

Table 3 Summary of the blood plasma coagulation cascade (CC) properties of the poly(alkyl methacrylate) coatings used in this study Polymer

Glass transition temperature (Tg ) (
C)

Onset time (min)

Fibrin deposition rate, (CCk), Hz/min

Coagulation time, t (min)

PIBMA PBMA PHMA PLMA POMA

66 17 6 70 100

2771 2475 1671 1871 2473

1371 1779 2771 3679 1173

94714 7273 3471 3271 73710

Shown is the lag phase time or time until onset (7standard error, SE), fibrin deposition rate (7SE) and coagulation time (7SE). The average of at least 2 measurements is presented.

and the production of fibrinogen starts. The general trend of the lag phase was the opposite of the IC results, i.e. the time until onset increased with decreased molecular mobility as can be seen in Fig. 6. It should be stressed that POMA did not follow this trend, which can be seen in Fig. 6 and in Table 3, where the results of all coatings are summarised. The coatings displayed a lag phase of 16–27 min before the surface deposition of fibrin started. This lag phase was somewhat longer for the poly(alkyl methacrylate) coatings than for a highly thrombogenic surface such as Titanium (Ti), to which the time until onset was determined to be around 8 min, Fig. 6. As expected, no change in frequency was observed during the first 60 min for a Heparinised surface, Fig. 6. The lag phase times measured in this study can be compared to the 20 min of lag phase that was measured when using polystyrene (PS) as substrate [15]. Furthermore, when measuring the contact activation of blood plasma Sa! nchez et al. measured an initial lag phase of 31 min with polytetrafluoroethylene (PTFE), and a significantly shorter lag phase of 6.2 min with poly(ethylene therephthalate) (PET or Dacron) [7]. It should be noted that a different experimental set-up was used to determine the lag phase in our study compared to the set-up used by Sa! nchez and co-workers. In another study, the lag phase of contact activation of plasma on contact with Gold (Au) was measured [37]. In that study the clot formation was stimulated by adding thromboplastin to the measurement chamber and significantly shorter lag phase times were measured (100–400 s) compared to the lag phase times measured in this study.

The fibrin deposition rate (CCk) decreased with decreased molecular mobility as can be seen in Fig. 6 and Table 3. Again POMA did not follow this trend. The poly(alkyl methacrylate) coatings displayed a fibrin deposition rate of 11–36 Hz min1. The deposition rate was comparable to that achieved with PS (15 Hz min1) [15]. As comparison, the deposition rate of a highly thrombogenic surface (Ti) was determined to 60 Hz min1. Finally, the coagulation time was determined, and as can be seen in Fig. 6 and Table 3, the coagulation time increased with decreased mobility. Note that POMA instead showed an increased coagulation time compared to the other polymers. In addition to the onset time, fibrin deposition rate and the coagulation time, the final level or the total shift of frequency differs between the polymers as can be seen in Fig. 6. It has been shown that the porosity, fibre dimension and architecture of fibrin gels formed in recalcified plasma are determined by the initial rate of fibrinogen activation [38]. A higher deposition rate and a shorter clotting time gives tighter gel architecture. As can be seen in Fig. 6, the poly(alkyl methacrylate) coatings form three different groups, which can be compared to the groups found when studying the IC activation. Thus, we have POMA in one group, PLMA, PHMA and PBMA in one group, and finally PIBMA in the last group. To be able to further understand the effect of surface mobility on the fibrin gel architecture other methods such as TM-AFM, scanning electron microscopy (SEM) or confocal microscopy should be used in combination with the ABBA methodology.

ARTICLE IN PRESS M. Berglin et al. / Biomaterials 25 (2004) 4581–4590

mobility. It seems that different aspects of the chemistry of the solid surface regulate activation of the IC and the CC. This is, to our knowledge, the first systematic study of the effect of polymer chain mobility on the IC and the CC activation. Furthermore, the Acoustic Blood Biointerface Analysis (ABBA) method based on a quartz crystal microbalance (QCM) was shown to be a very simple and fast methodological platform for the investigation of the surface induced IC and CC activation.

Fibrin deposition rate (∆ Hz/min)

50

40

PLMA

30

PHMA

20

4589

PBMA PIBMA POMA

10

Acknowledgements 0 0

5

10

15

20

25

30

Immune complement activity (∆ Hz)

Fig. 7. The fibrin deposition rate (Hz/min, 7SE) plotted as a function of immune complement activation (DHz, 7SE). Note the discrepancy of POMA compared to the other poly(alkyl methacrylate) coatings.

Financial support from Entific Medical Systems AB, . Goteborg, is gratefully acknowledged. The authors thank Andrea de Bejczy whose suggestions made significant improvements to this article.

References In general, we observed the opposite behaviour when comparing the IC activity and the CC properties. Hence, PIBMA that showed the highest IC activity showed a prolonged lag phase, a slower fibrin deposition rate and a prolonged coagulation time compared to, for example, PLMA. Plotting the fibrin deposition rate (CCk) against the IC activity for all poly(alkyl methacrylate) coatings can be seen in Fig. 7. Note the discrepancy of POMA in Fig. 7. Even though POMA did not follow the trend we can conclude that the molecular flexibility had a major effect on both IC and CC and it seems that different aspects of the chemistry of the solid surface regulates the activation of the IC and the CC. POMA showed both the lowest IC activation and the lowest fibrin deposition rate this result need some attention. As discussed above it has been shown that the relatively long alkyl chains in POMA are able to crystallise and that might affect the response. Composite surfaces with variable adhesiveness and surfaces with microstructures have shown interesting biological responses in several other studies [39–41]. But ultimately, the IC and CC response depends on more than just the molecular mobility as discussed above and the result of POMA verifies to the complexity of this problem.

4. Conclusions The results obtained in this study, using a series of poly(alkyl methacrylate) coatings, suggest that the surface induced immune complement (IC) activation is reduced with increased polymer chain mobility. In contrast, the surface induced plasma coagulation cascade (CC) showed no correlation with molecular

[1] Sefton MV, Sawyer A, Gorbet M, Black JP, Cheng E, Gemmell C, Pottinger-Cooper E. Does surface chemistry affect thrombogenicity of surface modified polymers? J Biomed Mater Res 2001;55(4):447–59. [2] Janatova J, Cheung AK, Parker CJ. Bio-medical polymers differ in their capacity to activate complement. Complement Inflammation 1991;8(2):61–9. [3] Tang LP, Liu L, Elwing HB. Complement activation and inflammation triggered by model biomaterial surfaces. J Biomed Mater Res 1998;41(2):333–40. [4] Luck M, Schroder W, Paulke BR, Blunk T, Muller RH. Complement activation by model drug carriers for intravenous application: determination by two-dimensional electrophoresis. Biomaterials 1999;20(21):2063–8. [5] Walivaara B, Aronsson BO, Rodahl M, Lausmaa J, Tengvall P. Titanium with different oxides—in-vitro studies of protein adsorption and contact activation. Biomaterials 1994;15(10): 827–34. [6] Rihova B. Biocompatibility of biomaterials: hemocompatibility, immunocompatibility and biocompatibility of solid polymeric materials and soluble targetable polymeric carriers. Adv Drug Delivery Rev 1996;21(2):157–76. [7] Sanchez J, Lundquist PB, Elgue G, Larsson R, Olsson P. Measuring the degree of plasma contact activation induced by artificial materials. Thromb Res 2002;105(5):407–12. [8] Lindhout T, Blezer R, Maassen C, Heijnen V, Reutelingsperger CPM. Platelet procoagulant surface as an essential parameter for the in-vitro evaluation of the blood compatibility of polymers. J Mater Sci Mater Med 1995;6(6):367–72. [9] Hong J, Andersson J, Ekdahl KN, Elgue G, Axen N, Larsson R, Nilsson B. Titanium is a highly thrombogenic biomaterial: possible implications for osteogenesis. Thromb Haemostasis 1999;82(1):58–64. [10] Burger R, Bader A, Kirschfink M, Rother U, Schrod L, Worner I, Zilow G. Functional analysis and quantification of the complement C3 derived anaphylatoxin C3a with a monoclonal antibody. Clin Exp Immunol 1987;68(3):703–11. [11] Black JP, Sefton MV. Complement activation by PVA as measured by ELIFA (enzyme-linked immunoflow assay) for SC5b-9. Biomaterials 2000;21(22):2287–94.

ARTICLE IN PRESS 4590

M. Berglin et al. / Biomaterials 25 (2004) 4581–4590

[12] Mollnes TE, Riesenfeld J, Garred P, Nordstrom E, Hogasen K, Fosse E, Gotze O, Harboe M. A new model for evaluation of biocompatibility: combined determination of neoepitopes in blood and on artificial surfaces demonstrates reduced complement activation by immobilization of heparin. Artif Organs 1995; 19(9):909–17. [13] Sellborn A, Andersson M, Fant C, Gretzer C, Elwing H. Methods for research on immune complement activation on modified sensor surfaces. Colloids Surf B Biointerfaces 2003;27(4):295–301. [14] Rodahl M, Hook F, Krozer A, Brzezinski P, Kasemo B. Quartzcrystal microbalance setup for frequency and q-factor measurements in gaseous and liquid environments. Rev Sci Instrum 1995;66(7):3924–30. [15] Andersson M, Sellborn A, Fant C, Gretzer C, Elwing H. Acoustics of blood plasma on solid surfaces. J Biomater Sci Polymer Edition 2002;13(8):907–17. [16] Garbassi F, Morra M, Occhiello E. Polymer surfaces, from physics to technology. New York: Wiley; 1994. [17] Wulf M, Grundke K, Kwok DY, Neumann AW. Influence of different alkyl side chains on solid surface tension of polymethacrylates. J Appl Polym Sci 2000;77(11):2493–504. [18] Rodahl M, Kasemo B. A simple setup to simultaneously measure the resonant frequency and the absolute dissipation factor of a quartz crystal microbalance. Rev Sci Instrum 1996;67(9):3238–41. [19] Saurbrey G. The use of quartzoscillators for weighing thin layers and for micoweighing. Physics 1959;155:206–22. [20] Tang LP, Eaton JW. Inflammatory responses to biomaterials. Am J Clin Pathol 1995;103(4):466–71. [21] Larsson R, Larm O, Olsson P. The search for thromboresistance using immobilized heparin. Annals New York Acad Sci 1987; 516:102–15. [22] Van Damme HS, Hogt AH, Feijen J. Surface mobility and structural transitions of poly(normal-alkyl methacrylates) probed by dynamic contact-angle measurements. J Colloid Interface Sci 1986;114(1):167–72. [23] Hook F, Kasemo B, Nylander T, Fant C, Sott K, Elwing H. Variations in coupled water, viscoelastic properties, and film thickness of a Mefp-1 protein film during adsorption and crosslinking: a quartz crystal microbalance with dissipation monitoring, ellipsometry, and surface plasmon resonance study. Anal Chem 2001;73(24):5796–804. [24] Mansfield KF, Theodorou DN. Molecular-dynamics simulation of a glassy polymer surface. Macromolecules 1991;24(23): 6283–94. [25] Norde W, Giacomelli CE. BSA structural changes during homomolecular exchange between the adsorbed and the dissolved states. J Biotechnol 2000;79(3):259–68. [26] Maste MCL, Norde W, Visser A. Adsorption-induced conformational changes in the serine proteinase savinase: a tryptophan fluorescence and circular dichroism study. J Colloid Interface Sci 1997;196(2):224–30.

[27] Elwing H, Nilsson B, Svensson KE, Askendahl A, Nilsson UR, Lundstrom I. Conformational-changes of a model protein (complement factor-Iii) adsorbed on hydrophilic and hydrophobic solid-surfaces. J Colloid Interface Sci 1988;125(1): 139–45. [28] Andersson J, Ekdahl KN, Larsson R, Nilsson UR, Nilsson B. C3 adsorbed to a polymer surface can form an initiating alternative pathway convertase. J Immunol 2002;168(11):5786–91. [29] Andrade JD, Gregonis DE, Smith LM. Physicochemical aspects of polymer surfaces. New York: Plenum Press; 1982. [30] Extrand CW, Kumagai Y. Contact angles and hysteresis on soft surfaces. J Colloid Interface Sci 1996;184(1):191–200. [31] Mosqueira VC, Legrand P, Gulik A, Bourdon O, Gref R, Labarre D, Barratt G. Relationship between complement activation, cellular uptake and surface physicochemical aspects of novel PEG-modified nanocapsules. Biomaterials 2001;22(22): 2967–79. [32] Vermette P, Meagher L. Interactions of phospholipid- and poly(ethylene glycol)-modified surfaces with biological systems: relation to physico-chemical properties and mechanisms. Colloids Surfaces B Biointerfaces 2003;28(2–3):153–98. [33] Alig I, Jarek M, Hellmann GP. Restricted segmental mobility in side-chain crystalline comblike polymers, studied by dielectric relaxation measurements. Macromolecules 1998;31(7):2245–51. [34] Mierzwa M, Floudas G, Stepanek P, Wegner G. Effect of pressure on the side-chain crystallization of poly(n-octadecyl methacrylate) studied by dielectric spectroscopy. Phys Rev B 2000;62(21):14012–9. [35] Biggs DL, Lengsfeld CS, Hybertson BM, Ng KY, Manning MC, Randolph TW. In vitro and in vivo evaluation of the effects of PLA microparticle crystallinity on cellular response. J Control Release 2003;92(1–2):147–61. [36] Kanazawa Y, Yoshida T, Kojima K. Structure–activity relationships in allergic contact dermatitis induced by methacrylates— studies of the influence of side-chain length of methacrylates. Contact Dermatitis 1999;40(1):19–23. [37] Vikinge TP, Hansson KM, Sandstrom P, Liedberg B, Lindahl TL, Lundstrom I, Tengvall P, Hook F. Comparison of surface plasmon resonance and quartz crystal microbalance in the study of whole blood and plasma coagulation. Biosensors Bioelectron 2000;15(11–12):605–13. [38] Blomback B, Carlsson K, Fatah K, Hessel B, Procyk R. Fibrin in human plasma—gel architectures governed by rate and nature of fibrinogen activation. Thromb Res 1994;76(5):501–2. [39] Harris A. Behaviour of cultured cells on substrata of variable adhesiveness. Exp Cell Res 1973;77:285–97. [40] Cheroudi B, Gould TR, Brunette DM. Effects of a grooved epoxy substratum on epithelial cell behaviour in vitro and in vivo. J Biomed Mater Res 1988;22:459–73. [41] Schmidt JA, Vonrecum AF. Macrophage response to microtextured silicone. Biomaterials 1992;13(15):1059–69.